Talaria

Minimal local tools for Intervals.icu.

Talaria uploads plain-text workouts and turns completed Intervals.icu activity data into an AI-coach-ready training summary. It runs locally on macOS, keeps API keys on your machine, and uses the Intervals.icu API directly.

What It Does

• Upload planned workouts from local .txt or .md files, using Intervals.icu workout syntax
• Parse one dated workout file, folders of files, or a single .md workout plan split by dated headings.
• Filter/sort workouts by sport, preview the folder tree, and open the import folder from the upload tab.
• Delete local workout files after upload if you no longer need them.
• Read completed activities, HR/load signals, wellness, intervals, streams, weather, athlete settings, gear, routes, calendar context, and curve context.
• Generate copyable coach summaries as reports.
• Save long reports as local Markdown files.
• Open a native app first, with an optional web UI button.

Requirements

• macOS
• Python 3 with Tkinter for the native app
• Intervals.icu API key

Start

Double-click Talaria.app.

If macOS is awkward after moving the folder, the fallback launchers still live inside the app package:

Talaria.app/Contents/Resources/app/launch_talaria_native.command
Talaria.app/Contents/Resources/app/launch_talaria_browser.command

Privacy

Talaria has no hosted backend. Your API key is stored locally in:

~/Library/Application Support/Talaria/intervals_config.json

Saved reports live inside:

Talaria.app/Contents/Resources/app/reports

Example workouts live inside:

Talaria.app/Contents/Resources/app/inputs/intervals_workouts/examples